The Obvious Switch

[straube]

Partner proposed using the Obvious Switch (Granovetter wrote about it) for defense, but I remember Rodwell saying that attitude signals should be about the suit being lead and not about some other. What are the experts doing now?

[dburn]

To the best of my knowledge, they are playing as Rodwell suggests (and as they have been playing for decades). That is: if they want partner to continue the suit led, they encourage it; if they want partner not to continue the suit led, they discourage it and hope that partner can work out what they want him to do instead.

 

They do a couple of other things also, but these may safely be dismissed as idiosyncrasies. A problem I faced only the day before yesterday concerned this hand:

 

♠J103 ♥J10765 ♦K ♣K1084

 

You pass at unfavourable vulnerability, LHO passes, partner opens 1♥ (in principle five, but may be four in third seat), RHO doubles. What call do you make? (2NT sound raise; any number of hearts pre-emptive raise; 3♣ clubs and hearts; no other artificial calls available).

 

You may wonder what this has to do with the question in the OP, but all will become clear. In the meantime, I thought it was an interesting bidding problem.

[bluecalm]

I think I would bid 2NT.

[jjbrr]

lol i read the OP and then the third message, but the second message was tl;dr. "I think I would bid 2NT" seemed like a pretty good response about Obvious Shift.

 

 

Anyway, I prefer attitude in the suit led.

[Lobowolf]

I wouldn't look at it as a pure dichotomy, though; part of answering the question "Do I want my partner to continue the suit led?" involves an awareness of what partner will likely do if he DOESN'T continue. The Granovetters took the principle to the nth degree, and created a scheme of rules addressing situations in which the "obvious shift" isn't so obvious, but the basic obvious shift principle preceded the publication of A Switch in Time.

[Phil]

♠J103 ♥J10765 ♦K ♣K1084

 

You pass at unfavourable vulnerability, LHO passes, partner opens 1♥ (in principle five, but may be four in third seat), RHO doubles. What call do you make? (2NT sound raise; any number of hearts pre-emptive raise; 3♣ clubs and hearts; no other artificial calls available).

I can't improvise with 3N and expect partner not to understand? If not, then 4♥.

 

I didn't think Meckwell used OS by the book, but they did have some similar method but I do not know what it is specifically.

 

I'm guessing Dburn is having us lead the ♥J against 4♠ and we have to work out the switch based on pard's card.

[mike777]

so we are suggested

 

not OS 100% by the book but very close for starters?

 

I am lucky if i can convince exp pards to play 50% sigh.......they know 100% of whatever....I dont.....

 

typical is that at trick one...........often....suit pref not att...sigh........I try.....

[awm]

I've played obvious shift in several partnerships and been very happy with the method.

 

I think it's fairly common at the expert level to encourage a suit when you don't really want partner to shift away from an honor in some side suit; in other words, even though you have nothing much in the suit lead, you think continuing it is a better defensive strategy than the alternatives.

 

Perhaps the point is that some hands call for a passive defense, and encouraging the opening lead is more likely to help partner find such defense. Discouraging thus says not only "I don't want you to continue the suit you lead" but also "please play something else at your first opportunity."

 

With that said, not many expert pairs have a formal set of agreements about which suit is the "obvious shift" that a discouraging signal requests partner to switch to. However, usually this suit is sufficiently "obvious" that it's not much of an issue.

[Codo]

Experts do not need OS, they can work it out for themselves when to signal for continuation and when for a shift.

 

But for lesser players OS is a great tool. Simply because it make you think like an expert on any given defensive hand.

You do not give a card for the lead suit in isolation, but you are forced to look at the complete hand.

 

This should be done always, but millions of players fail to do so every single day.

So OS will improve the defence of most players dramatically. Not because it is a superior system (it is not) but because it teaches to think about the complete hand.

 

You do already? Always? No need to switch in time.

[Free]

One of my partners insisted to play OS. I've been playing OS with him for more than a year now, and I haven't encountered any problems so far. Instead of thinking like an expert and failing miserably, we prefer a strict rule which works in almost every situation. In fact, we've had some great defenses, even when we ask for a switch which is not the OS (throwing an honour card under partner's Ace or dummy's higher honour)! Without the clear signal, it would've been too risky to switch to that suit.

[gwnn]

I would bid 4♥. I do not have much to contribute to the issue of OS since none of my partners have proposed it to me and neither have I to them, but I felt the need to clarify that it's not just bluecalm who is interested in what happens next.

[Codo]

2 Nt and what happens next?

[bluecalm]

Btw, for people who have some experience actually playing obvious switch (I don't have any, I've skimmed through Granovetter's book): can you please give more agreements you have/had ?

Do you play obvious switch in every situation ?

If not what are the situations which you play attitude in lead suit (or count) ?

 

The method I prefer playing is simple attituide and if that's not relevant then count and if that's not relevant (AK sec in dummy, 3 stoppers etc) then suit preference but then we give suit preference in trumps and other suits played by declarer if they don't require count (partner holding up with an honour etc.). I would like to hear more about relative merits of those methods.

[pooltuna]

To the best of my knowledge, they are playing as Rodwell suggests (and as they have been playing for decades). That is: if they want partner to continue the suit led, they encourage it; if they want partner not to continue the suit led, they discourage it and hope that partner can work out what they want him to do instead.

 

They do a couple of other things also, but these may safely be dismissed as idiosyncrasies. A problem I faced only the day before yesterday concerned this hand:

 

♠J103 ♥J10765 ♦K ♣K1084

 

You pass at unfavourable vulnerability, LHO passes, partner opens 1♥ (in principle five, but may be four in third seat), RHO doubles. What call do you make? (2NT sound raise; any number of hearts pre-emptive raise; 3♣ clubs and hearts; no other artificial calls available).

 

You may wonder what this has to do with the question in the OP, but all will become clear. In the meantime, I thought it was an interesting bidding problem.

I would try 4♥

[Phil]

I should have mentioned that I have played it for about 10 years in one partnership. He loves it; I like it.

 

It gives an immense amount of information away to declarer, which is the main drawback. But as Free states, you can practically defend double dummy in a lot of cases.

[mikeh]

I played OS in what has been my most successful partnership until we stopped playing in 2000. It made defence so easy: going back to normal defence, even with expert partners, was like someone turning out the lights in a room at dusk: I could still see, but nothing was as clear as it had been.

 

When we first started using it, and getting good results, we went overboard, like many a convert. We then found that expert declarers were picking us apart, because we were sending out far too many accurate signals.

 

So we limited our OS....always on the opening lead, and never after about trick 4 or 5....and the extent to which we used OS at tricks 2-5 depended on what was going on....generally, if OS applied (ie no other signal took priority) we did OS for the first 3-4 tricks. And suit preference in trump throughout.

 

Edit: re free's modest remark that he likes OS because it is strict rules and one doesn't have to think like an expert: OS is not a substitute for judgement: it is a method that allows the defenders to exchange more information than can most, and to then apply their judgement with the benefit of the information. It can turn an average defender into an apparent expert defender, and make an expert seem world class. The key is that signaller uses judgement to decide what signal to give and partner uses judgement to decide what to do with the information conveyed.

[straube]

The response so far is more positive than I would have thought. I wonder what Fred Gitelman thinks of it (hint).

[bluecalm]

So we limited our OS....always on the opening lead, and never after about trick 4 or 5....and the extent to which we used OS at tricks 2-5 depended on what was going on....generally, if OS applied (ie no other signal took priority) we did OS for the first 3-4 tricks. And suit preference in trump throughout.

 

How signalling OS works after you already gave one such signal ?

So let's say on opening leads you played low to show that you can tolerate obvious switch, which Axx suit in dummy.

Declarer won 1st trick and played other suit in which you have another chance to signal. You play low. Does it mean you have KQ/KJ/QJ now or something else about your hand ?

[Phil]

If you play OS by the book, subsequent cards after T1 by both defenders are meant to convey suit preference. This is primarily valuable by opening leader.

 

I also think Smith Echo is very important by responder when playing OS.

 

There is judgment involved. When the board has KJT and that's the OS, don't automatically signal encouragement with the Q, even though that's conventionally what you're supposed to do.

[kgr]

Do you play obvious switch in every situation ?

If not what are the situations which you play attitude in lead suit (or count) ?

We play OS in every situation (Also when trump is lead, or with a singleton in dummy).

The next signal(s) when declarer leads are Lavinthal until we told enough.

(When it probably costs a trick if partner continues the opening suit and we don't have the OS suit then we can lie about the OS).

[awm]

Sometimes it's not clear whether the best signal at trick one is attitude, suit-preference, or count. Most top-flight partnerships have at least some agreements about these situations (i.e. count to ace lead at five-level or above, suit preference when dummy has singleton in the suit lead, etc) but there are still many situations which are unclear. I know that some pairs like to "try to signal what partner needs to know", which is great when you are on the same wavelength and terrible otherwise (I have seen both happen at the table to well-established pairs), and other pairs like to have as many rules as possible and follow them (guaranteeing that both know what the signal means, but implying that sometimes the signal is not the "most useful" information available).

 

One nice aspect of obvious shift is that it combines the attitude and suit-preference signals at trick one. This means the only remaining decision is "do we need to give count here" which hopefully is easier to determine than having to decide between all three options. There are however a few situations where you do need to give count at trick one (most commonly when partner leads the ace from ace-king in his long suit and dummy has several small cards, and we need to know how many tricks to cash before declarer can ruff).

 

The Switch in Time book suggests giving suit preference signals usually during the hand (unless count is clearly necessary). I think this is a good method (have played it a lot in my strongest regular partnership) for several reasons, one of which is that count signals are very commonly useful to declarer. Obviously you can give false count, but this means you have to constantly think about it (should I give true count or false count? who if anyone will it help? do I believe partner's count signal?) and also reduces the proportion of true signals during the hand. Suit preference signals can of course help declarer too, but I find that most of the time the information declarer gets out of these is less useful (i.e. "you are going down") and that it helps partner more. It's also a little easier to think about false-carding these, because it's more closely related to what you have to actually figure out in defending the hand than count signals (usually) are.

 

I also think that if you're playing "frequent suit preference signals" as the book suggests, you don't really need to play smith echo. Having seen a very established partnership screw up smith echo in GNT open just yesterday, this seems like an appealing thing to me.

[Siegmund]

I am a big fan of having an Obvious Shift agreement. It is, as a previous poster pointed out, not so much a "different kind of signal" as thinking about what partner will most likely switch to if I discourage the suit he has led. It spectacularly improved my defense as soon as I started thinking about what the obvious shift was.

 

That said, I did NOT so much care for the REST of Granovetter's book ("all suit preference all the time.") In the regular partnership where we adopted Obvious Shift, we kept all of our existing agreements about signal priorities (basically "attitude first, unless obvious or irrelevant, then count, unless obvious or irrelevant, then suit preference" ), and just added to it that all attitude signals were made mode precise via OS Principle. The only other change we made to our system was using an early trump echo to ask for the non-obvious shift.

 

Granovetter claimed, incidentally, that something like OS was expert standard, and I tend to believe him (all the defense books talk about how you must sometimes give false preference to avoid a disastrous shift, etc, they just don't talk much about how to evaluate the relative importance of the suit you're signaling for vs. the potential switch.)

[jdonn]

I have really never seen how obvious shift differs from simple attitude signals. Partner signals encouragement anyway if he can't stand the shift he expects you to make if discouraged. As far as I can tell the only relevance is in defining which suit is the obvious shift suit in close situations. But if it's a close situation I don't want to rule out the second-most-obvious shift suit in any case, so on that basis it seems a little unsound to me.

[awm]

I have really never seen how obvious shift differs from simple attitude signals. Partner signals encouragement anyway if he can't stand the shift he expects you to make if discouraged. As far as I can tell the only relevance is in defining which suit is the obvious shift suit in close situations. But if it's a close situation I don't want to rule out the second-most-obvious shift suit in any case, so on that basis it seems a little unsound to me.

Well there are several differences. Here are some examples:

 

(1) Opening lead is a suit that we're highly unlikely to want to continue (i.e. dummy has a singleton or ace-king-tight or something). Playing obvious shift, partner will discourage if he wants a shift to a particular side suit (the "obvious shift") and encourage if he either has no interest in either likely shift or wants a shift to the other side suit. If he dumps an unnecessary honor card it is a clear signal for the non-obvious shift side suit. Playing standard methods, I think the usual thing to do in this situation is to give a suit-preference signal. Of course this seems to accomplish the same thing (although which card is played to signal which suit will be different); however obvious shift helps if it is not 100% clear that continuing the lead suit is non-sensical (i.e. there is more difference between the "normal signal" and the "dummy has a stiff" signal playing standard carding).

 

(2) There has been an uninformative auction and dummy has two pretty similar looking suits aside from the one lead (i.e. say 1♠-3♠-4♠, partner leads a club, dummy hits with king-third in both reds). Playing standard methods, if we discourage the opening lead it will be clear that partner should switch to a red suit, but not which one. Some people try to solve this by playing suit preference with their spot cards (i.e. playing udca, small club is encouraging, medium club is for diamonds, big club is for hearts) but trying to squish three possible meanings into one signal often makes the spots much harder to read. Playing obvious shift, one of the suits is identified as the "obvious shift suit" (normally diamonds, the tie-breaker is lower suit) and a discouraging trick one signal says "switch to this suit." This makes it a lot easier to give a clear signal for diamonds and somewhat harder to signal for hearts (sometimes we encourage when we want a heart shift, although if we can dump an honor card on the first club it is a clearer signal).

 

(3) In general, it occasionally comes up that you want partner to make a weird shift that he's unlikely to think of on his own. Like say you have a void in dummy's very strong suit; it can be very hard to get partner to switch to that suit since an encouraging signal often gets partner to continue the suit lead, and a discouraging signal gets the "normal shift" from partner to try to establish side suit tricks before dummy's long suit runs. Obvious shift has this special agreement about dumping honors to get the "non-obvious" shift. Of course, sometimes people play that anyway but at least here there is a clear-cut rule about which shift is "non-obvious" (i.e. say they have an auction to 4♥, partner leads the ace of diamonds from ace-king and dummy hits with five clubs to the KQJ, a stiff spade, three small diamonds, and some trumps -- how do you get partner to switch to a club at trick two without cashing a second diamond?)

 

(4) The follow-up strategy of giving suit-preference to declarer's leads rather than count is very different from standard signals. It occasionally bites you (when you need count and pd doesn't signal it) but it also can be extremely helpful.

[dake50]

I play FARTP: ForceAttackRuffTrumpPassive in that order.

So when I discourage it discourages the FARTP aspect started.

Let partner 'see' what change I hope for -not some conventionalized Obvious Switch.